An LLM-Augmented Multi-Modal Forensic Reasoning Framework for Detecting Tampered and AI-Generated Digital Images and Digital Documents
Year: 2026
Type: article
Abstract: Tampering detection of images, and of AI-generated (GAN/diffusion-synthesized) images in particular, and forensic inconsistencies of PDF, OOXML, and legacy binary documents are important issues related to the evaluation of integrity, authenticity and trustworthiness of digital evidence. Traditional single-signal detectors provide a label lacking interpretable evidence trail and thus are useless when justification of the verdict is needed. This paper introduces a desktop forensic analysis software called ForensiX AI that performs analysis of image and document forensics.
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7237361973
